How every number is made

We follow the standard of Wikipedia’s “List of Jewish Nobel laureates”: people who are Jewish or have at least one Jewish parent. That list names 223 individuals from 1905 to 2025, and we matched every one to the Nobel Foundation’s official record.

Each profile records the actual basis (both parents, father, mother, conversion and so on) and the laureate’s own relationship to Jewish identity, with sources. Some laureates were raised in another faith or did not identify as Jewish; they are included under the parentage standard, and their profiles say so plainly. The list also includes a few people with more distant Jewish ancestry, such as a Jewish grandparent. They are flagged as disputed, their profiles explain the evidence, and the results without them are shown below.

We never claim that being Jewish explains anyone’s achievement. The standard only defines the group being counted. Explanations for the pattern are historical and cultural, and they are discussed on the context page.

Laureate means an individual person. The 28 organizations that have won the Peace Prize are excluded from every count. There are 990 individual laureates; 5 of them won twice, so there are 995 individual awards. None of the double winners is on the Jewish list, so Jewish laureates and Jewish awards are both 223.

Categories are the six Nobel prizes, with the Sveriges Riksbank Prize in Economic Sciences (from 1969) included as “economics”. Prizes are dated by award year. Birth country uses today’s borders, as the Nobel Foundation records them.

The live results first, then how each one is computed. The number in each label points to its formula below.

All formulas are implemented once, in a pure TypeScript statistics package with unit tests against hand-computed values, and every page uses that package.

Each profile may include quantified claims about real-world outcomes the laureate’s work contributed to. Every claim states the whole outcome as a low–high range cumulative from a start year to 2025, the share credited to the laureate, whether it is direct or ripple, how it was derived (a published total or explicit arithmetic from cited inputs), and a confidence rating.

credited = whole outcome × attribution share

• No double counting. Laureates who contributed to the same outcome share one outcome ID. The site takes the median of their ranges and adds their shares, capped at 100%.
• Direct and ripple. Direct means the laureate’s own discovery and its immediate uses. Ripple means downstream effects, which are modeled and usually low confidence. Ripple is off by default.
• Harms count too. Deaths and injuries caused (poison gas, weapons) use the same method and appear alongside the benefits.
• Over time. Each outcome is spread from its start year to 2025 along an adoption curve (logistic, linear, exponential or one-time).
• No forced numbers. Literature, most peace prizes and pure theory often carry no number at all, only a description.

Each profile follows a written research contract. It must cite at least three sources, including the Nobel Foundation, with every factual statement traceable to a cited source. Writers paraphrase rather than copy, quote the laureate at most once, never quote poetry or lyrics, and state harms plainly.

A separate, adversarial fact-check then re-opens the sources, checks dates against the official record, tests every number and the Jewish background claim, removes anything it cannot support, and records what it changed. Profiles show the date of that check.

Profiles are drafted with AI research assistance, working from the cited sources under these rules. The fact-check is a separate, adversarial AI pass: it did not write the draft, and it re-opens the sources rather than trusting them. Before launch, an editor will review the voice and spot-check profiles against their sources; that human review has not happened yet. Every claim links to its source so that anyone can verify it.

Status: 223 of 223 profiles published.

5 laureates are flagged as disputed. We count them by default, and show here how the results change without them.

• Gustav Hertz (1925)
• Hermann J. Muller (1946)
• Igor Y. Tamm (1958)
• Aage N. Bohr (1975)
• Roger Penrose (2020)

Separately, 50 laureates are recorded as refugees or exiles who fled persecution themselves or as children.

• Who counts as Jewish is a definition, not a fact of nature. Other standards (self-identification only, or two Jewish parents) would give different counts.
• Population estimates before 1950 are uncertain, and Jewish population estimates vary by definition (core versus enlarged). The time-weighted ratio inherits that uncertainty.
• The Nobel Prize is one narrow measure of contribution. It has well-documented biases, including against women and against work done outside Europe and North America.
• A ratio describes a gap; it does not explain it. Correlation with a group is not a cause.
• Impact estimates are ranges built from published figures and explicit assumptions. They are meant to be defensible orders of magnitude, not precise counts.

If you find an error, we want to fix it. Every profile has a “Suggest a correction” link, or you can write to corrections@nobeljews.com. Please include a source. Corrections are made in the public data files and noted on the profile.

All data behind the site is free to reuse with attribution.

• laureates.csvOne row per laureate: prize, year, birth country, flags
• laureates-index.jsonLaureate cards with taglines and impact domains
• stats-inputs.jsonAward years, category counts, population series, country counts
• impact-claims.jsonEvery quantified impact claim with its reasoning
• context.jsonPopulation, timeline, factors, comparisons and baselines
• anomalies.jsonThe comparison cases